Popular text-to-image (T2I) systems are trained on web-scraped data, which is heavily Amero and Euro-centric, underrepresenting the cultures of the Global South. To analyze these biases, we introduce CuRe, a novel and scalable benchmarking and scoring suite for cultural representativeness that leverages the marginal utility of attribute specification to T2I systems as a proxy for human judgments. Our CuRe benchmark dataset has a novel categorical hierarchy built from the crowdsourced Wikimedia knowledge graph, with 300 cultural artifacts across 32 cultural subcategories grouped into six broad cultural axes (food, art, fashion, architecture, celebrations, and people). Our dataset's categorical hierarchy enables CuRe scorers to evaluate T2I systems by analyzing their response to increasing the informativeness of text conditioning, enabling fine-grained cultural comparisons. We empirically observe much stronger correlations of our class of scorers to human judgments of perceptual similarity, image-text alignment, and cultural diversity across image encoders (SigLIP 2, AIMV2 and DINOv2), multimodal language models (OpenCLIP, SigLIP 2, Gemini 2.0 Flash) and state-of-the-art text-to-image systems, including three variants of Stable Diffusion (1.5, XL, 3.5 Large), FLUX.1 [dev], Ideogram 2.0, and DALL-E 3. The code and dataset is open-sourced and available at https://aniketrege.github.io/cure/.
论文检索
输入标题、作者或关键词，从 9,256 篇学术成果中精准定位
While diffusion models show promising results in image editing given a target prompt, achieving both prompt fidelity and background preservation remains difficult. Recent works have introduced score distillation techniques that leverage the rich generative prior of text-to-image diffusion models to solve this task without additional fine-tuning. However, these methods often struggle with tasks such as object insertion. Our investigation of these failures reveals significant variations in gradient magnitude and spatial distribution, making hyperparameter tuning highly input-specific or unsuccessful. To address this, we propose two simple yet effective modifications: attention-based spatial regularization and gradient filtering-normalization, both aimed at reducing these variations during gradient updates. Experimental results show our method outperforms state-of-the-art score distillation techniques in prompt fidelity, improving successful edits while preserving the background. Users also preferred our method over state-of-the-art techniques across three metrics, and by 58-64% overall.
Diffusion models have fundamentally transformed the field of generative models, making the assessment of similarity between customized model outputs and reference inputs critically important. However, traditional perceptual similarity metrics operate primarily at the pixel and patch levels, comparing low-level colors and textures but failing to capture mid-level similarities and differences in image layout, object pose, and semantic content. Contrastive learning-based CLIP and self-supervised learning-based DINO are often used to measure semantic similarity, but they highly compress image features, inadequately assessing appearance details. This paper is the first to discover that pretrained diffusion models can be utilized for measuring visual similarity and introduces the DiffSim method, addressing the limitations of traditional metrics in capturing perceptual consistency in custom generation tasks. By aligning features in the attention layers of the denoising U-Net, DiffSim evaluates both appearance and style similarity, showing superior alignment with human visual preferences. Additionally, we introduce the Sref and IP benchmarks to evaluate visual similarity at the level of style and instance, respectively. Comprehensive evaluations across multiple benchmarks demonstrate that DiffSim achieves state-of-the-art performance, providing a robust tool for measuring visual coherence in generative models.
Domain Generalization aims to develop models that can generalize to novel and unseen data distributions. In this work, we study how model architectures and pre-training objectives impact feature richness and propose a method to effectively leverage them for domain generalization. Specifically, given a pre-trained feature space, we first discover latent domain structures, referred to as pseudo-domains, that capture domain-specific variations in an unsupervised manner. Next, we augment existing classifiers with these complementary pseudo-domain representations making them more amenable to diverse unseen test domains. We analyze how different pre-training feature spaces differ in the domain-specific variances they capture. Our empirical studies reveal that features from diffusion models excel at separating domains in the absence of explicit domain labels and capture nuanced domain-specific information. On 5 datasets, we show that our very simple framework improves generalization to unseen domains by a maximum test accuracy improvement of over 4% compared to the standard baseline Empirical Risk Minimization (ERM). Crucially, our method outperforms most algorithms that access domain labels during training. Code is available at: https://xthomasbu.github.io/GUIDE/.
Diffusion Transformer (DiT)-based generation models have achieved remarkable success in video generation. However, their inherent computational demands pose significant efficiency challenges. In this paper, we exploit the inherent temporal non-uniformity of real-world videos, and observe that videos exhibit dynamic information density, with high-motion segments demanding greater detail preservation than static scenes. Inspired by this temporal non-uniformity, we propose DLFR-Gen, a training-free approach for Diffusion-based Video Generation with Dynamic Latent Frame Rate. DLFR-Gen adaptively adjusts the number of elements in latent space based on the motion frequency of the latent space content, using fewer tokens for low-frequency segments while preserving detail in high-frequency segments. Specifically, our key contributions are: (1) A dynamic frame rate scheduler for DiT video generation that adaptively assigns frame rates for video segments. (2) A novel latent-space frame merging method to align latent representations with their denoised counterparts before merging those redundant in low-resolution space. (3) A preference analysis of Rotary Positional Embeddings (RoPE) across DiT layers, informing a tailored RoPE strategy optimized for semantic and local information capture. Experiments show that DLFR-Gen can achieve a speedup up to 3x for video generation with minimal quality degradation.
Recent 3D facial reconstruction methods have made significant progress in shape estimation, but high-fidelity unbiased facial albedo estimation remains challenging. Existing methods rely on expensive light-stage captured data, and while they have made progress in either high-fidelity reconstruction or unbiased skin tone estimation, no work has yet achieved optimal results in both aspects simultaneously. In this paper, we present a novel high-fidelity unbiased facial diffuse albedo reconstruction method, HUST, which recovers the diffuse albedo map directly from a single image without the need for captured data. Our key insight is that the albedo map is the illumination-invariant texture map, which enables us to use inexpensive texture data for diffuse albedo estimation by eliminating illumination. To achieve this, we collect large-scale high-resolution facial images and train a VQGAN model in the image space. To adapt the pre-trained VQGAN model for UV texture generation, we fine-tune the encoder by using limited UV textures and our high-resolution faces under adversarial supervision in both image and latent space. Finally, we train a cross-attention module and utilize group identity loss for the domain adaptation from texture to albedo. Extensive experiments demonstrate that HUST can predict high-fidelity facial albedos for in-the-wild images. On the FAIR benchmark, HUST achieves the lowest average ITA error (11.20) and bias score (1.58), demonstrating superior accuracy and robust fairness across the entire spectrum of human skin tones. Our code, models, and training data will be made publicly available to facilitate future research.
GestureHYDRA: Semantic Co-speech Gesture Synthesis via Hybrid Modality Diffusion Transformer and Cascaded-Synchronized Retrieval-Augmented Generation
PDF ↗While increasing attention has been paid to co-speech gesture synthesis, most previous works neglect to investigate hand gestures with explicit and essential semantics. In this paper, we study co-speech gesture generation with an emphasis on specific hand gesture activation, which can deliver more instructional information than common body movements. To achieve this, we first build a high-quality dataset of 3D human body movements including a set of semantically explicit hand gestures that are commonly used by live streamers. Then we present a hybrid-modality gesture generation system GestureHYDRA built upon a hybrid-modality diffusion transformer architecture with novelly designed motion-style injective transformer layers, which enables advanced gesture modeling ability and versatile gesture operations. To guarantee these specific hand gestures can be activated, we introduce a cascaded retrieval-augmented generation strategy built upon a semantic gesture repository annotated for each subject and an adaptive audio-gesture synchronization mechanism, which substantially improves semantic gesture activation and production efficiency. Quantitative and qualitative experiments demonstrate that our proposed approach achieves superior performance over all the counterparts.
We present REM, a framework for segmenting a wide range of concepts in video that can be described through natural language. Our method leverages the universal visual-language mapping learned by video diffusion models on Internet-scale data by fine-tuning them on small-scale Referring Object Segmentation datasets. Our key insight is to preserve the entirety of the generative model's architecture by shifting its objective from predicting noise to predicting mask latents. The resulting model can accurately segment rare and unseen objects, despite only being trained on a limited set of categories. Additionally, it can effortlessly generalize to non-object dynamic concepts, such as smoke or raindrops, as demonstrated in our new benchmark for Referring Video Process Segmentation (Ref-VPS). REM performs on par with the state-of-the-art on in-domain datasets, like Ref-DAVIS, while outperforming them by up to 12 IoU points out-of-domain, leveraging the power of generative pre-training. We also show that advancements in video generation directly improve segmentation.
3D style transfer refers to the artistic stylization of 3D assets based on reference style images. Recently, 3DGS-based stylization methods have drawn considerable attention, primarily due to their markedly enhanced training and rendering speeds. However, a vital challenge for 3D style transfer is to strike a balance between the content and the patterns and colors of the style. Although the existing methods strive to achieve relatively balanced outcomes, the fixed-output paradigm struggles to adapt to the diverse content-style balance requirements from different users. In this work, we introduce a creative intensity-tunable 3D style transfer paradigm, dubbed Tune-Your-Style, which allows users to flexibly adjust the style intensity injected into the scene to match their desired content-style balance, thus enhancing the customizability of 3D style transfer. To achieve this goal, we first introduce Gaussian neurons to explicitly model the style intensity and parameterize a learnable style tuner to achieve intensity-tunable style injection. To facilitate the learning of tunable stylization, we further propose the tunable stylization guidance, which obtains multi-view consistent stylized views from diffusion models through cross-view style alignment, and then employs a two-stage optimization strategy to provide stable and efficient guidance by modulating the balance between full-style guidance from the stylized views and zero-style guidance from the initial rendering. Extensive experiments demonstrate that our method not only delivers visually appealing results, but also exhibits flexible customizability for 3D style transfer. Project page is available at https://zhao-yian.github.io/TuneStyle.
Generative videos have the potential to revolutionize game development by autonomously creating new content. In this paper, we present GameFactory, a framework for action-controlled scene-generalizable game video generation. We first address the fundamental challenge of action controllability by introducing GF-Minecraft, an action-annotated game video dataset without human bias, and developing an action control module that enables precise control over both keyboard and mouse inputs. We further extend to support autoregressive generation for unlimited-length interactive videos.More importantly, GameFactory tackles the critical challenge of scene-generalizable action control, which most existing methods fail to address. To enable the creation of entirely new and diverse games beyond fixed styles and scenes, we leverage the open-domain generative priors from pre-trained video diffusion models. To bridge the domain gap between open-domain priors and small-scale game datasets, we propose a multi-phase training strategy with a domain adapter that decouples game style learning from action control. This decoupling ensures that action control learning is no longer bound to specific game styles, thereby achieving scene-generalizable action control. Experimental results demonstrate that GameFactory effectively generates open-domain action-controllable game videos, representing a significant step forward in AI-driven game generation.
In this paper, we investigate when and how visual representations learned by two different generative models diverge from each other. Specifically, given two text-to-image models, our goal is to discover visual attributes that appear in images generated by one model but not the other, along with the types of prompts that trigger these attribute differences. For example, 'flames' might appear in one model's outputs when given prompts expressing strong emotions, while the other model does not produce this attribute given the same prompts. We introduce CompCon (Comparing Concepts), an evolutionary search algorithm that discovers visual attributes more prevalent in one model's output than the other, and uncovers the prompt concepts linked to these visual differences. To evaluate CompCon's ability to find diverging representations, we create an automated data generation pipeline to produce ID^2, a dataset of 60 input-dependent differences, and compare our approach to several LLM- and VLM-powered baselines. Finally, we use CompCon to compare popular text-to-image models, finding divergent representations such as how PixArt depicts prompts mentioning loneliness with wet streets and Stable Diffusion 3.5 depicts African American people in media professions.
With its extensive applications, Foreground Conditioned Out-painting (FCO) has attracted considerable attention in the research field. Through the utilization of text-driven FCO, users are enabled to generate diverse backgrounds for a given foreground by adjusting the text prompt, which considerably enhances the efficiency in fields like e-commerce. Since the foreground is fixed in FCO, a key concern is whether the generated background can match the foreground well to achieve a coherent composition. However, most existing methods are lacking in this regard. Artifacts and incorrect interactions are common defects in synthesized images. This issue is linked to the influence of the initial noise in the sampling process. As the initial noise is sampled independently, it's highly likely that the implied image composition will conflict with the given foreground. In this paper, a novel Initialization Policy Model (IPM) is proposed to address this problem. Its function is to replace the early denoising steps and directly predict the intermediate state that is conducive to the reasonable image composition. Since the IPM is designed to take only the foreground image and the text prompt as inputs, it isolates the impact of the initial noise. The subsequently proposed training paradigm that combines inversion-derived label supervision and diffusion reward supervision further ensures the efficient training of the IPM. The evaluation is conducted using the task-specific OpenImage-FCO dataset developed by us. The results verify that the introduction of the IPM can significantly improve the composition of the synthesized images and achieve advanced performance in the FCO task.
We introduce StreamDiffusion, a real-time diffusion pipeline designed for streaming image generation. Existing diffusion models are adept at creating images from text or image prompts, yet they often fall short in real-time interaction. This limitation becomes particularly evident in scenarios involving continuous input, such as augmented/virtual reality, video game graphics rendering, live video streaming, and broadcasting, where high throughput is imperative. StreamDiffusion tackles this challenge through a novel pipeline-level system design. It employs unique strategies like batching the denoising process (Stream Batch), residual classifier-free guidance (R-CFG), and stochastic similarity filtering (SSF). Additionally, it seamlessly integrates advanced acceleration technologies for maximum efficiency. Specifically, Stream Batch reformulates the denoising process by eliminating the traditional wait-and-execute approach and utilizing a batching denoising approach, facilitating fluid and high-throughput streams. This results in 1.5x higher throughput compared to the conventional sequential denoising approach. R-CFG significantly addresses inefficiencies caused by repetitive computations during denoising. It optimizes the process to require minimal or no additional computations, leading to speed improvements of up to 2.05x compared to previous classifier-free methods. Besides, our stochastic similarity filtering dramatically lowers GPU activation frequency by halting computations for static image flows, achieving a remarkable reduction in computational consumption--2.39 times on an RTX 3060 GPU and 1.99 times on an RTX 4090 GPU, respectively. The synergy of our proposed strategies with established acceleration technologies enables image generation to reach speeds of up to 91.07 fps on a single RTX 4090 GPU, outperforming the throughput of AutoPipeline, developed by Diffusers, by more than 59.56x.
Diffusion probabilistic models have shown significant progress in video generation; however, their computational efficiency is limited by the large number of sampling steps required. Reducing sampling steps often compromises video quality or generation diversity. In this work, we introduce a distillation method that combines variational score distillation and consistency distillation to achieve few-step video generation, maintaining both high quality and diversity. We also propose a latent reward model fine-tuning approach to further enhance video generation performance according to any specified reward metric. This approach reduces memory usage and does not require the reward to be differentiable. Our method demonstrates state-of-the-art performance in few-step generation for 10-second videos (128 frames at 12 FPS). The distilled student model achieves a score of 82.57 on VBench, surpassing the teacher model as well as baseline models Gen-3, T2V-Turbo, and Kling. One-step distillation accelerates the teacher model's diffusion sampling by up to 278.6 times, enabling near real-time generation. Human evaluations further validate the superior performance of our 4-step student models compared to teacher model using 50-step DDIM sampling.
FreeDNA: Endowing Domain Adaptation of Diffusion-Based Dense Prediction with Training-Free Domain Noise Alignment
PDF ↗Domain Adaptation (DA) for dense prediction tasks is an important topic, which enhances the dense prediction model's performance when tested on its unseen domain. Recently, with the development of Diffusion-based Dense Prediction (DDP) models, the exploration of DA designs tailored to this framework is worth exploring, since the diffusion model is effective in modeling the distribution transformation that comprises domain information. In this work, we propose a training-free mechanism for DDP frameworks, endowing them with DA capabilities. Our motivation arises from the observation that the exposure bias (e.g., noise statistics bias) in diffusion brings domain shift, and different domains in conditions of DDP models can also be effectively captured by the noise prediction statistics. Based on this, we propose a training-free Domain Noise Alignment (DNA) approach, which alleviates the variations of noise statistics to domain changes during the diffusion sampling process, thereby achieving domain adaptation. Specifically, when the source domain is available, we directly adopt the DNA method to achieve domain adaptation by aligning the noise statistics of the target domain with those of the source domain. For the more challenging source-free DA, inspired by the observation that regions closer to the source domain exhibit higher confidence meeting variations of sampling noise, we utilize the statistics from the high-confidence regions progressively to guide the noise statistic adjustment during the sampling process. Notably, our method demonstrates the effectiveness of enhancing the DA capability of DDP models across four common dense prediction tasks. Code is available at \href https://github.com/xuhang07/FreeDNA https://github.com/xuhang07/FreeDNA .
Recent interactive matting methods have demonstrated satisfactory performance in capturing the primary regions of objects, but they fall short in extracting fine-grained details in edge regions. Diffusion models trained on billions of image-text pairs, demonstrate exceptional capability in modeling highly complex data distributions and synthesizing realistic texture details, while exhibiting robust text-driven interaction capabilities, making them an attractive solution for interactive matting. To this end, we propose SDMatte, a diffusion-driven interactive matting model, with three key contributions. First, we exploit the powerful priors of the pre-trained U-Net within diffusion models and transform the text-driven interaction mechanism into a visual prompt-driven interaction mechanism to enable interactive matting. Second, we integrate coordinate embeddings of visual prompts and opacity embeddings of objects into U-Net, enhancing SDMatte's sensitivity to spatial position information and opacity information. Third, we propose a masked self-attention mechanism and a visual prompt-driven interaction mechanism that enable the model to focus on areas specified by visual prompts, leading to better performance. Extensive experiments on multiple datasets demonstrate the superior performance of our method, validating its effectiveness in interactive matting. Code will be made publicly available.
EfficientMT: Efficient Temporal Adaptation for Motion Transfer in Text-to-Video Diffusion Models
PDF ↗The progress on generative models has led to significant advances on text-to-video (T2V) generation, yet the motion controllability of generated videos remains limited. Existing motion transfer approaches explored the motion representations of reference videos to guide generation. Nevertheless, these methods typically rely on sample-specific optimization frameworks, resulting in high computational burdens. In this paper, we propose EfficientMT, a novel and efficient end-to-end framework for video motion transfer. By leveraging a small set of synthetic paired motion transfer samples, EfficientMT effectively adapts a pretrained T2V model into a general motion transfer framework that can accurately capture and reproduce diverse motion patterns. Specifically, we repurpose the backbone of the T2V model to extract temporal information from reference videos, and further propose a scaler module to distill motion-related information. Subsequently, we introduce a temporal integration mechanism that seamlessly incorporates reference motion features into the video generation process. After training on our self-collected synthetic paired samples, EfficientMT enables general video motion transfer without requiring test-time optimization. Extensive experiments demonstrate that our EfficientMT outperforms existing methods in efficiency while maintaining flexible motion controllability. Our code will be available at https://github.com/PrototypeNx/EfficientMT.
DC-ControlNet: Decoupling Inter- and Intra-Element Conditions in Image Generation with Diffusion Models
PDF ↗In this paper, we introduce DC (Decouple)-ControlNet, a highly flexible and precisely controllable framework for multi-condition image generation. The core idea behind DC-ControlNet is to decouple control conditions, transforming global control into a hierarchical system that integrates distinct elements, contents, and layouts. This enables users to mix these individual conditions with greater flexibility, leading to more efficient and accurate image generation control. Previous ControlNet-based models rely solely on global conditions, which affect the entire image and lack the ability of element- or region-specific control. This limitation reduces flexibility and can cause condition misunderstandings in multi-conditional image generation. To address these challenges, we propose both intra-element and inter-element Controllers in DC-ControlNet. The Intra-Element Controller handles different types of control signals within individual elements, accurately describing the content and layout characteristics of the object. For interactions between elements, we introduce the Inter-Element Controller, which accurately handles multi-element interactions and occlusion based on user-defined relationships. Extensive evaluations show that DC-ControlNet significantly outperforms existing ControlNet models and Layout-to-Image generative models in terms of control flexibility and precision in multi-condition control.
Safeguarding Vision-Language Models: Mitigating Vulnerabilities to Gaussian Noise in Perturbation-based Attacks
PDF ↗Vision-Language Models (VLMs) extend the capabilities of Large Language Models (LLMs) by incorporating visual information, yet they remain vulnerable to jailbreak attacks, especially when processing noisy or corrupted images. Although existing VLMs adopt security measures during training to mitigate such attacks, vulnerabilities associated with noise-augmented visual inputs are overlooked. In this work, we identify that missing noise-augmented training causes critical security gaps: many VLMs are susceptible to even simple perturbations such as Gaussian noise. To address this challenge, we propose Robust-VLGuard, a multimodal safety dataset with aligned / misaligned image-text pairs, combined with noise-augmented fine-tuning that reduces attack success rates while preserving functionality of VLM. For stronger optimization-based visual perturbation attacks, we propose DiffPure-VLM, leveraging diffusion models to convert adversarial perturbations into Gaussian-like noise, which can be defended by VLMs with noise-augmented safety fine-tuning. Experimental results demonstrate that the distribution-shifting property of diffusion model aligns well with our fine-tuned VLMs, significantly mitigating adversarial perturbations across varying intensities. The dataset and code are available at https://github.com/JarvisUSTC/DiffPure-RobustVLM.
Mixture-of-Experts-based (MoE-based) diffusion models demonstrate remarkable scalability in high-fidelity image generation, yet their reliance on expert parallelism introduces critical communication bottlenecks. State-of-the-art methods alleviate such overhead in parallel diffusion inference through computation-communication overlapping, termed displaced parallelism. However, we identify that these techniques induce severe staleness --- the usage of outdated activations from previous timesteps that significantly degrades quality, especially in expert-parallel scenarios. We tackle this fundamental tension and propose DICE, a staleness-centric optimization framework with a three-fold approach: (1) Interweaved Parallelism introduces staggered pipelines, effectively halving step-level staleness for free; (2) Selective Synchronization operates at layer-level and protects layers vulnerable to staled activations; and (3) Conditional Communication, a token-level, training-free method that dynamically adjusts communication frequency based on token importance. Together, these strategies effectively reduce staleness, achieving 1.26x speedup with minimal quality degradation. Empirical results establish DICE as an effective and scalable solution. Our code is available at https://github.com/Cobalt-27/DICE